Summary: This week’s podcast was recorded on the road a few weeks ago in Silicon Valley. Chris dropped into the <a href="https://www.datrium.com/">Datrium</a> offices in Sunnyvale to talk about DVX, Datrium’s open converged platform. Taking part in the podcast is <a href="https://twitter.com/sazzala">Sazzala Reddy</a>, CTO and co-founder and <a href="https://twitter.com/tushar_agr">Tushar Agrawal</a>, Director of Products.<br> Datrium DVX is an extension of HCI that scales performance and capacity separately. Persistent storage is deployed on a shared array, with each compute host using local cache in the form of flash storage. You can find out more on a <a href="https://blog.architecting.it/modern-storage-architectures-datrium/">recent blog post</a> covering DVX.<br> The conversation covers an explanation on why traditional HCI has issues scaling and how Datrium has solved some of these challenges. This leads on to a discussion on the features of DVX releases, including DVX 4.0, which offers backup to public cloud.<br> During the podcast Sazzala mentions Little’s Law (<a href="https://en.wikipedia.org/wiki/Little%27s_law">details here</a>). He also discusses <a href="https://en.wikipedia.org/wiki/Log-structured_file_system">Log Structured File Systems</a> and VMware founder <a href="https://en.wikipedia.org/wiki/Mendel_Rosenblum">Mendel Rosenblum</a>. A copy of his paper can be found <a href="https://people.eecs.berkeley.edu/~brewer/cs262/LFS.pdf">here</a>. Also, here’s the original Google <a href="https://research.google.com/archive/mapreduce-osdi04.pdf">MapReduce</a> paper.<br> Elapsed Time: 00:32:30<br> Timeline<br> <br> * 00:00:00 – Intro<br> * 00:01:00 – The challenges of the data centre<br> * 00:02:00 – The heritage of knowledge on data deduplication<br> * 00:03:00 – Understanding the Datrium architecture<br> * 00:06:00 – What are the issues with traditional HCI?<br> * 00:08:00 – Splitting storage performance and capacity<br> * 00:10:00 – What improvements are customers seeing?<br> * 00:12:00 – Log Structured File System implementation<br> * 00:14:00 – How does failover/recovery work?<br> * 00:15:00 – How does the solution address some of the backup issues?<br> * 00:19:00 – Deep Dive – Global Deduplication, compression, data integrity<br> * 00:24:30 – Deployment models<br> * 00:27:00 – DVX Version 4.0 and a recap of the previous releases<br> * 00:30:30 – What is Cloud DVX?<br> * 00:31:30 – Wrap Up<br> <br> <br> Copyright (c) 2016-2018 Storage Unpacked. No reproduction or re-use without permission.<br>
